Speech-to-text (STT) and text-to-speech (TTS) systems enable microphone-equipped computing devices to interpret speech and text. These systems provide an alternative to conventional human-to-computer input devices, such as keyboards or keypads. Many telecommunications devices are equipped with automated speech recognition (ASR) technology that detects the presence of discrete speech such as a spoken nametag or control vocabulary in the form of numerals, keywords, or commands.
However, some ASR systems fail to recognize speech from a user when it includes unusual pronunciations. This is especially true for words or names with unusual pronunciations. For example, a name might be affected by regional accents, dialects, slang, silent letters, unfamiliar consonant blends, ethnicity of the name, or any other variations in name pronunciation. Even if the ASR system correctly identifies the speech, the system TTS output to the user may differ from what is expected.